Skip to content
View PratyushJha254's full-sized avatar

Block or report PratyushJha254

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
PratyushJha254/README.md

Hi there 👋 I'm Pratyush Jha

AI Research Engineer @ Ola Krutrim · B.Tech (Engineering Physics), IIT Roorkee (2024)

I enjoy using mathematical insight to design better models and to connect theoretical ideas with practical systems. Apart from this, working on the intersection of AI and Physics is especially exciting to me.


🔎 Quick summary

  • Current Role: AI Research Engineer, Ola Krutrim (2024–Present)
  • Education: B.Tech in Engineering Physics, IIT Roorkee (2024)
  • Interests: Model interpretability, LLM reasoning, reinforcement learning for alignment and reasoning, and physics-inspired generative modeling

🌐 Research Highlights

  • AI Research Engineer @ Ola Krutrim (2024–Present): Worked on improving LLM reasoning and reducing hallucinations using preference optimization techniques, while also building scalable AI systems and services—from model deployment pipelines to real‑time analytics and developer SDKs.

  • B.Tech Thesis (2023–24): Explored deep learning-based approaches to simulate phase diagrams in condensed matter physics. Compared fine-tuning and transfer learning strategies to study the models' generalization abilities.

  • MITACS Globalink Intern @ McGill University (Summer 2023): Modeled CR‑39 neutron detectors with TOPAS-based Monte Carlo simulations to study detector response to high-energy and thermal neutrons. 🔗 Lab Profile


🔍 Research themes

  • Improving intuition in LLMs using Reinforcement learning
  • Reinforcement Learning for reasoning alignment and preference tuning
  • Physics-inspired generative modeling — diffusion models for condensed-matter systems

🧪 Projects & technical skills

  • Preference tuning, reasoning alignment experiments, and evaluation frameworks
  • End-to-end AI pipelines: data preparation, training, deployment, real-time analytics
  • Production & research tooling: Python, C++, Hugging Face, FastAPI, Flask, Docker, Redis, NVIDIA DeepStream

🤝 Open to collaborations & roles

I’m extremely happy to work on research projects and open to exploring new domains in AI. If you have a project, internship, or role where I can contribute (or learn), let’s talk.


📫 Contact

Popular repositories Loading

  1. RL-Mario-Agent RL-Mario-Agent Public

    Jupyter Notebook

  2. Lab-Based-Project Lab-Based-Project Public

    This is with respect to the Lab Based Project under Professor Peayush Choubey.

    Jupyter Notebook

  3. Feature-Based-Dynamic-Pricing-Paper-Implementation-Cohen-et-al.-2020- Feature-Based-Dynamic-Pricing-Paper-Implementation-Cohen-et-al.-2020- Public

    Jupyter Notebook

  4. BTech-Project BTech-Project Public

    Jupyter Notebook 1

  5. Nucleosynthesis-Processes-in-Stars-Term-Paper Nucleosynthesis-Processes-in-Stars-Term-Paper Public

  6. MultiPDF-Insight-Engine-RAG-Application-using-Google-Gemini MultiPDF-Insight-Engine-RAG-Application-using-Google-Gemini Public

    Jupyter Notebook